A two-stage robust optimization model for the planning capacity of the new power system "source-storage" of new power system is established, taking into account the uncertainty of power output and load demand. Using even theory and Karush Kuhn Tucker conditions (KKT), the robust optimization problem with min-max-min structure is decomposed into two stages: the main problem and sub problems for solution. The main problem is the minimization optimization problem structure, which adopts a relaxed form and considers the carbon emission requirements of new power system and the constraints of wind and solar power curtailment. The optimal values of the installed capacity of various power sources are determined, and the upper bound of the optimal solution of the original problem is obtained. The sub-problem is a maximum-minimum optimization problem structure containing nonlinear variables. It is modeled using dual theory and KKT conditions, and the Big-M method is used to handle nonlinear constraints. A box type uncertain set is used to describe the uncertain scenarios of wind and solar power output. Based on the optimization results of the first stage, various worst-case scenarios are determined. The simulation results of three scenarios in the example validate the feasibility and adaptability of the proposed model and method.